Alexa support/API that allows me to build it?

Any chance of Alexa support? As far as I can tell, that wouldn't require text between plays. Alternately, can I retrieve a secret but non-authenticated URL to the mp3 to stream? That way I could tell the echo to start that stream.

Comments

  • KoenKoen Administrator
    edited November 2020

    The developer documentation is here: https://asti.ga/developers

    Every parameter works both as part of a GET request or POST request (in the body), so you could construct a "secret" URL like https://play.asti.ga/rest/stream.view?id=1234 (with of course all the mandatory parameters added as well, see the "required parameters" section). The ID can be found using one of the other methods (like search).

    Regarding Alexa support, you're not the first to ask it:

    I will have to look into that. I currently do not own an Echo (although Amazon does now sell the in the Netherlands as well).

  • KoenKoen Administrator

    The community search uses Google's custom site search engine, and it seems to be thoroughly confused by the fact that I added a sitemap. So confused it seems to have forgotten about everything. I have removed that sitemap, and hopefully it will resolve itself automatically in a few days.

  • Just dropping by to show my support. I frequently use Alexa support to play music, especially in the car with my Echo Auto, and it boggles the mind that there’s no configurable subsonic player skill that can be connected to Astiga. (Maybe there’s a technical or policy reason why such an open-ended skill either doesn’t exist or at least isn’t made public?) I’d love to see a skill developed at some point. I use a mirror of my library stored on iBroadcast for this purpose, but having literally no other option to play my own music using Alexa is bothersome.

  • KoenKoen Administrator

    @Cralex There are some reasons. One of them is that you need a publicly available URL for your service. In the case of Subsonic, this is generally not the case, as your Subsonic may be hosted at contoso.org, and someone else's Subsonic is hosted at fabrikam.com. If you don't have that, then publishing your skill/action is difficult, as Amazon/Google generally require that, unless you require everyone to clone the skill, and adapt it to their situation, but that is tricky again, because you cannot publish it (imagine if everyone could ask "Hey Google, talk to Bob's Subsonic", except for 1000+ users). You can get around it by proxying all requests through your own server, and have everyone link their personal server at one central place, but that means hosting a server, which is already quite a hurdle. It is one of the reasons why Home Assistant, an open source home automation toolkit, offers Google Assistant support via their special cloud offering: https://www.home-assistant.io/integrations/google_assistant/

    So, that's the first reason why having skills/actions for self-hosted software is difficult. Now, Astiga is not self-hosted, so that is not directly a problem. However, both Google and Amazon have their own partners when it comes to music streaming. Everyone can create a skill or action for music streaming, but it is not as seamless as the built-in ones. For one, playing something through the Astiga action on Google Assistant means saying something like "Hey Google, ask Astiga to play Never Gonna Give You Up". If you then say "Hey Google, pause", and then ask "Hey Google, what's the weather like?", you'd expect to hear "sunny with a chance of rain" (or whatever your weather is). Instead, you will hear "Sorry, I don't understand", because you are still talking to Astiga. You then first need to exit out of Astiga before asking Google a question again. It also means I have to reinvent the wheel for all things you could ask a music player, like shuffle, repeat, etc.

    So that's some background as to why. I have considered making a skill for the Echo that supports both Astiga and Subsonic, but the second issue is the main one holding me back. I am currently trying to find a way to smoothen the Google experience, by not necessarily having a separate action invocation, but that is something that will probably take months rather than days.

Sign In or Register to comment.